Modulating the Light Switch by 3MLCT-3ππ* State Interconversion

Nov 23, 2010 - The origin of this light switch behavior involves two competing 3MLCT states .... interacting 3MLCT excited states, a highly emissive â...
0 downloads 0 Views 2MB Size
Download PDF
Inorg. Chem. 2010, 49, 11333–11345 11333 DOI: 10.1021/ic1011617

Modulating the Light Switch by 3MLCT-3ππ* State Interconversion Brigitte R. Spencer, Brian J. Kraft, Chris G. Hughes, Maren Pink, and Jeffrey M. Zaleski* Department of Chemistry and Molecular Structure Center, Indiana University, Bloomington, Indiana 47405, United States Received June 9, 2010

The spectroscopic, electronic, and DNA-binding characteristics of two novel ruthenium complexes based on the dialkynyl ligands 2,3-bis(phenylethynyl)-1,4,8,9-tetraaza-triphenylene (bptt, 1) and 2,3-bis(4-tert-butyl-phenylethynyl)-1,4,8,9-tetraaza-triphenylene (tbptt, 2) have been investigated. Electronic structure calculations of bptt reveal that the frontier molecular orbitals are localized on the pyrazine-dialkynyl portion of the free ligand, a property that is reflected in a red shift of the lowest energy electronic transition (1: λmax = 393 nm) upon substitution at the terminal phenyl groups (2: λmax = 398 nm). Upon coordination to ruthenium, the low-energy ligand-centered transitions of 1 and 2 are retained, and metal-to-ligand charge transfer transitions (MLCT) centered at λmax = 450 nm are observed for [Ru(phen)2bptt]2þ(3) and [Ru(phen)2tbptt]2þ(4). The photophysical characteristics of 3 and 4 in ethanol closely parallel those observed for [Ru(bpy)3]2þ and [Ru(phen)3]2þ, indicating that the MLCT excited state is primarily localized within the [Ru(phen)3]2þ manifold of 3 and 4, and is only sparingly affected by the extended conjugation of the bptt framework. In an aqueous environment, 3 and 4 possess notably small luminescence quantum yields (3: φH2O = 0.005, 4: φH2O = 0.011) and biexponential decay kinetics (3: τ1 = 40 ns, τ2 = 230 ns; 4: τ1 ∼ 26 ns, τ2 = 150 ns). Addition of CT-DNA to an aqueous solution of 3 causes a significant increase in the luminescence quantum yield (φDNA = 0.045), while the quantum yield of 4 is relatively unaffected (φDNA = 0.013). The differential behavior demonstrates that tert-butyl substitution on the terminal phenyl groups inhibits the ability of 4 to intercalate with DNA. Such changes in intrinsic luminescence demonstrate that 3 binds to DNA via intercalation (Kb = 3.3  104 M-1). The origin of this light switch behavior involves two competing 3MLCT states similar to that of the extensively studied light switch molecule [Ru(phen)2dppz]2þ. The solvent- and temperature-dependence of the luminescence of 3 reveal that the extended ligand aromaticity lowers the energy of the 3ππ* excited state into competition with the emitting 3MLCT state. Interconversion between these two states plays a significant role in the observed photophysics and is responsible for the dual emission in aqueous environments.

Introduction Genotoxic agents including carcinogens or radiation exposure can cause DNA defects such as base pair mismatches and abasic sites that can lead to permanent mutations and disruption of transcription if not properly repaired.1-4 If the genes that code for the repair machinery are themselves damaged and can no longer carry out their critical function, mutations can develop and accumulate leading to diseases such as cancer.4-6 Therefore, a great deal of interest lies in metallo-insertors and metallo-intercalators that can interact intimately with DNA base pairs upon binding, and therefore *To whom correspondence should be addressed. E-mail: zaleski@ indiana.edu. (1) Modrich, P. Annu. Rev. Genet. 1991, 25, 229–253. (2) Kolodner, R. Genes Dev. 1996, 10, 1433–1442. (3) Syvaenen, A.-C. Nat. Rev. Genet. 2001, 2, 930–942. (4) David, S. S.; O’Shea, V. L.; Kundu, S. Nature 2007, 447, 941–950. (5) Kolodner, R. D. Trends Biochem. Sci. 1995, 20, 397–401. (6) Arzimanoglou, I. I.; Gilbert, F.; Barber, H. R. Cancer 1998, 82, 1808– 1820. (7) Zeglis, B. M.; Pierre, V. C.; Barton, J. K. Chem. Commun. 2007, 4565– 4579.

r 2010 American Chemical Society

can be structurally tuned to target these specific DNA defects or even particular DNA sequences.7 Metallo-intercalators (e.g., [Ru(phen)2 dppz]2þ, [Ru(phen)3 ]2þ, [Cu(bpod)2 ]2þ, and [Pt(terpy)Cl]þ)8-14 bind between adjacent base pairs accompanied by a slight unwinding of the DNA helix for steric accommodation. In the case of insertion, the ligand is generally more sterically hindered, so to bind, a base pair must be ejected from the interior of the DNA helix. Therefore, metallo-insertors such as [Rh(bpy)2(chrysi)]3þ (8) Howe-Grant, M.; Wu, K. C.; Bauer, W. R.; Lippard, S. J. Biochemistry 1976, 15, 4339–4346. (9) Friedman, A. E.; Chambron, J. C.; Sauvage, J. P.; Turro, N. J.; Barton, J. K. J. Am. Chem. Soc. 1990, 112, 4960–4862. (10) Veal, J. M.; Rill, R. L. Biochemistry 1988, 27, 1822–1827. (11) Veal, J. M.; Rill, R. L. Biochemistry 1989, 28, 3243–3250. (12) Veal, J. M.; Rill, R. L. Biochemistry 1991, 30, 1132–1140. (13) Gorner, H.; Tossi, A. B.; Stradowski, C.; Schulte-Frohlinde, D. J. Photochem. Photobiol. B 1988, 2, 67–89. (14) Benites, P. J.; Holmberg, R. C.; Rawat, D. S.; Kraft, B. J.; Klein, L. J.; Peters, D. G.; Thorp, H. H.; Zaleski, J. M. J. Am. Chem. Soc. 2003, 125, 6434–6446. (15) Zeglis, B. M.; Boland, J. A.; Barton, J. K. J. Am. Chem. Soc. 2008, 130, 7530–7531.

Published on Web 11/23/2010

pubs.acs.org/IC

11334 Inorganic Chemistry, Vol. 49, No. 24, 2010 and [Rh(bpy)2(phzi)]3þ preferentially bind to sites of weaker base pair interactions such as mismatches or abasic sites.7,15-19 For example, the DNA-binding affinity of [Δ-Rh(bpy)2(chrysi)]3þ is 1000-fold greater for mismatched sites than for normal DNA base pairs.15,17,19 Analogously, [Ru(phen)2dppz]2þ preferentially undergoes metallo-insertion instead of intercalative binding upon encountering DNA mismatches and abasic sites.20 Additionally, sequence specific binding can be achieved by designing ancillary ligand targeting that maximize intermolecular interactions such as H-bonding and van der Waals contacts with the DNA sequence of interest.21,22 The ability to confidently rely on luminescent probes for DNA structural interrogation has only come through extensive characterization of the DNA binding modes and binding specificity of these types of transition metal complexes using X-ray crystallography, NMR spectroscopy, DNA footprinting and cleavage gel assays, and competitive binding displacement assays.7,15-17,19,20,23-36 Metal-based probes typically exhibit strong, MLCT-based molecular light switch emission, where a large increase in luminescence quantum yield is detectable upon DNA binding. The most widely studied light switch compound, [Ru(phen)2dppz]2þ, displays a 104-fold increase in luminescence intensity when bound to DNA in aqueous solution.9 Extensive photophysical studies have shown that the unusual binary luminescent behavior of [Ru(phen)2dppz]2þ in the presence and absence of DNA is the result of two interacting 3MLCT excited states, a highly emissive “light” state where the charge-transfer generated negative charge is localized on the phenanthroline components of the ligands (3MLCTphen), as well as a second, much more weakly emissive (16) Zeglis, B. M.; Boland, J. A.; Barton, J. K. Biochemistry 2009, 48, 839–849. (17) Zeglis, B. M.; Pierre, V. C.; Kaiser, J. T.; Barton, J. K. Biochemistry 2009, 48, 4247–4253. (18) Junicke, H.; Hart, J. R.; Kisko, J.; Glebov, O.; Kirsch, I. R.; Barton, J. K. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 3737–3742. (19) Zeglis, B. M.; Barton, J. K. Nat. Protoc. 2007, 2, 357–371. (20) Lim, M. H.; Song, H.; Olmon, E. D.; Dervan, E. E.; Barton, J. K. Inorg. Chem. 2009, 48, 5392–5397. (21) Sitlani, A.; Barton, J. K. Biochemistry 1994, 33, 12100–12108. (22) Sitlani, A.; Dupureur, C. M.; Barton, J. K. J. Am. Chem. Soc. 1993, 115, 12589–12590. (23) Kielkopf, C. L.; Erkkila, K. E.; Hudson, B. P.; Barton, J. K.; Rees, D. C. Nat. Struct. Biol. 2000, 7, 117–121. (24) Pierre, V. C.; Kaiser, J. T.; Barton, J. K. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 429–434. (25) Hudson, B. P.; Barton, J. K. J. Am. Chem. Soc. 1998, 120, 6877–6888. (26) Hudson, B. P.; Dupureur, C. M.; Barton, J. K. J. Am. Chem. Soc. 1995, 117, 9379–9380. (27) Rueba, E.; Hart, J. R.; Barton, J. K. Inorg. Chem. 2004, 43, 4570– 4578. (28) Zhen, Q. X.; Ye, B. H.; Liu, J. G.; Zhang, Q. L.; Ji, L. N.; Wang, L. Inorg. Chim. Acta 2000, 303, 141–147. (29) Zhen, Q. X.; Ye, B. H.; Zhang, Q. L.; Liu, J. G.; Li, H.; Ji, L. N.; Wang, L. J. Inorg. Biochem. 1999, 76, 47–53. (30) Chen, W.; Turro, C.; Friedman, L. A.; Barton, J. K.; Turro, N. J. J. Phys. Chem. B 1997, 101, 6995–7000. (31) Coates, C. G.; Callaghan, P.; McGarvey, J. J.; Kelly, J. M.; Jacquet, L.; Kirsch-De Mesmaeker, A. J. Mol. Struct. 2001, 598, 15–25. (32) Delaney, S.; Pascaly, M.; Bhattacharya, P. K.; Han, K.; Barton, J. K. Inorg. Chem. 2002, 41, 1966–1974. (33) Hiort, C.; Lincoln, P.; Norden, B. J. Am. Chem. Soc. 1993, 115, 3448– 3454. (34) Holmlin, R. E.; Stemp, E. D. A.; Barton, J. K. Inorg. Chem. 1998, 37, 29–34. (35) Yam, V. W.-W. V.; Lo, K. K.-W.; Cheung, K.-K.; Kong, R. Y.-C. J. Chem. Soc., Dalton Trans. 1997, 2067–2072. (36) Zhang, Q.-L.; Liu, J.-H.; Liu, J.-Z.; Zhang, P.-X.; Ren, X.-Z.; Liu, Y.; Huang, Y.; Ji, L.-N. J. Inorg. Biochem. 2004, 98, 1405–1412.

Spencer et al. “dark” state where the negative charge is localized on the phenazine portion of the dppz ligand (3MLCTphzn).9,30,34,37-49 The lower energy 3MLCTphzn is stabilized and deactivated in the presence of solvents that can H-bond to the periphery phenazine nitrogens resulting in quenching of the emissive 3 MLCTphen state. DNA intercalation of the dppz ligand leads to an increase in solvent shielding of the phenazine nitrogens, which reduces the non-radiative decay rate and increases the emissive quantum yield of the complex. Since [Ru(phen)2dppz]2þ has a very short lifetime (τ ∼ 3 ps) and marginal emission quantum yield (φ ∼ 10-7) in aqueous solution, the turn-on effect is significant.45 However, even DNA solvent shielding does not result in an extensive lifetime (τDNA ∼ 75, 250 ns) or an impressively high emission quantum yield (φDNA ∼ 0.02) as is typical of commercial dyebased diagnostics.9 One possible approach for enhancing the emissive lifetime and quantum yield is to perturb the excited state dynamics by mixing a low-lying 3ππ* ligand-based state with the emissive 3 MLCT state.50 Since these ligand states typically have much longer lifetimes, the luminescence characteristics of the complex can be tuned depending upon the thermodynamics of the 3 MLCT manifold and the intervening 3ππ* state. Within this theme, the excited state dynamics of some Re(I) dppz complexes do indeed exhibit perturbed kinetics because of rapid interconversion between the emissive 3MLCT and the ligandlocalized 3ππ* state.51,52 In contrast to Ru(II)-dppz complexes where the 3ππ* ligand-based state lies to higher energy than the 3MLCT manifold and plays no role in the observed photophysics,38 perturbations to the classic 3MLCT luminescence from the 3ππ* state are observed in fac-[Re(CO)3(dppz)L] complexes.51,52 The relative populations of the three states and the rate of interconversion between them are dependent on the structure and environment of the complex. (37) Brennaman, M. K.; Alstrum-Acevedo, J. H.; Fleming, C. N.; Jang, P.; Meyer, T. J.; Papanikolas, J. M. J. Am. Chem. Soc. 2002, 124, 15094– 15098. (38) Brennaman, M. K.; Meyer, T. J.; Papanikolas, J. M. J. Phys. Chem. A 2004, 108, 9938–9944. (39) Coates, C. G.; Callaghan, P. L.; McGarvey, J. J.; Kelly, J. M.; Kruger, P. E.; Higgins, M. E. J. Raman Spectrosc. 2000, 31, 283–288. (40) Coates, C. G.; McGarvey, J. J.; Callaghan, P. L.; Coletti, M.; Hamilton, J. G. J. Phys. Chem. B 2001, 105, 730–735. (41) Coates, C. G.; Olofsson, J.; Coletti, M.; McGarvey, J. J.; Oenfelt, B.; Lincoln, P.; Norden, B.; Tuite, E.; Matousek, P.; Parker, A. W. J. Phys. Chem. B 2001, 105, 12653–12664. (42) Nair, R. B.; Cullum, B. M.; Murphy, C. J. Inorg. Chem. 1997, 36, 962–965. (43) Nair, R. B.; Murphy, C. J. J. Inorg. Biochem. 1998, 69, 129–133. (44) Olofsson, J.; Onfelt, B.; Lincoln, P.; Norden, B.; Matousek, P.; Parker, A. W.; Tuite, E. J. Inorg. Biochem. 2002, 91, 286–297. (45) Olson, E. J. C.; Hu, D.; Hoermann, A.; Jonkman, A. M.; Arkin, M. R.; Stemp, E. D. A.; Barton, J. K.; Barbara, P. F. J. Am. Chem. Soc. 1997, 119, 11458–11467. (46) Onfelt, B.; Lincoln, P.; Norden, B.; Baskin, J. S.; Zewail, A. H. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 5708–5713. (47) Fees, J.; Kaim, W.; Moscherosch, M.; Matheis, W.; Klima, J.; Krejcik, M.; Zalis, S. Inorg. Chem. 1993, 32, 166–174. (48) Browne, W. R.; McGarvey, J. J. Coord. Chem. Rev. 2006, 250, 1696– 1709. (49) McGarvey, J. J.; Callaghan, P.; Coates, C. G.; Schoonover, J. R.; Kelly, J. M.; Jacquet, L.; Gordon, K. C. J. Phys. Chem. B 1998, 102, 5941– 5942. (50) Wang, X.-Y.; Del Guerzo, A.; Schmehl, R. H. J. Photochem. Photobiol., C 2004, 5, 55–77. (51) Kuimova, M. K.; Alsindi, W. Z.; Blake, A. J.; Davies, E. S.; Lampus, D. J.; Matousek, P.; McMaster, J.; Parker, A. W.; Towrie, M.; Sun, X.-Z.; Wilson, C.; George, M. W. Inorg. Chem. 2008, 47, 9857–9869. (52) Schoonover, J. R.; Bates, W. D.; Meyer, T. J. Inorg. Chem. 1995, 34, 6421–6422.

Article

As expected for these environment-sensitive molecules, the observed photophysics are variable upon ligand substitution, substituent modification of the dppz ligand, changes in solvent dielectric constant, or changes in temperature.51,52 Our interests in metal complex photophysics stem from the ability to employ extended π-conjugation53-55 or metalligand charge transfer excitation to promote ligand-based diradical generation for nuclease applications.14,56-58 In related work, we have developed the enediyne-chelated complex, [Cu(bpod)2]2þ (bpod = cis-1,8-bis(pyridin-3-oxy)oct4-ene-2,6-diyne), that can intercalate DNA and upon photoexcitation, performs DNA strand scission by C-40 H-atom abstraction.14 Unfortunately, DNA binding of this complex and related structures has been difficult to characterize since their spectroscopic signatures are only very modestly perturbed by interactions with DNA. The desire for compounds that act as self-reporters for DNA binding led us to develop enediyne analogues of [Ru(phen)2dppz]2þ based on the alkynylated ligands 2,3-bis(phenylethynyl)-1,4,8,9-tetraazatriphenylene (bptt, 1) and 2,3-bis(4-tert-butyl-phenylethynyl)1,4,8,9-tetraaza-triphenylene (tbptt, 2). [Ru(phen)2(bptt)]2þ exhibits molecular light switch behavior upon binding to DNA while the sterically hindered [Ru(phen)2(tbptt)]2þ complex is unable to intercalate into the DNA helix. More subtly, the extended π-conjugation of 1 relative to dppz lowers the energy of the 3ππ* excited state leading to participation of this state in the observed photophysical behavior, similar to that for Re(I) dppz complexes and unlike that for [Ru(phen)2dppz]2þ.

Experimental Section Materials. All air sensitive manipulations were performed under nitrogen using Schlenk and drybox techniques. Tetrahydrofuran was dried by distillation over sodium/benzophenone. All chemicals used in the syntheses and characterizations were of the highest purity available from Aldrich and Fluka and were used as received. Ru(phen)2Cl259 and 5,6-diamino-1,10-phenanthroline60,61 were prepared according to literature methods. The dione, (53) Chandra, T.; Kraft, B. J.; Huffman, J. C.; Zaleski, J. M. Inorg. Chem. 2003, 42, 5158–5172. (54) Nath, M.; Pink, M.; Zaleski, J. M. J. Am. Chem. Soc. 2005, 127, 478–479. (55) Koepke, T.; Pink, M.; Zaleski, J. M. Org. Biomol. Chem. 2006, 4, 4059–4062. (56) Kraft, B. J.; Eppley, H. J.; Huffman, J. C.; Zaleski, J. M. J. Am. Chem. Soc. 2002, 124, 272–280. (57) Kraft, B. J.; Zaleski, J. M. New J. Chem. 2001, 25, 1281–1289. (58) Maurer, T. D.; Kraft, B. J.; Lato, S. M.; Zaleski, J. M.; Ellington, A. D. Chem. Commun. 2000, 69–70. (59) Sullivan, B. P.; Salmon, D. J.; Meyer, T. J. Inorg. Chem. 1978, 17, 3334–3341. (60) Bodige, S.; MacDonnell, F. M. Tetrahedron Lett. 1997, 38, 8159– 8160. (61) Yamada, M.; Tanaka, Y.; Yoshimoto, Y.; Kuroda, S.; Shimao, I. Bull. Chem. Soc. Jpn. 1992, 65, 1006–1011. (62) Faust, R.; Weber, C.; Fiandanese, V.; Marchese, G.; Punzi, A. Tetrahedron 1997, 53, 14655–14670. (63) Choy, N.; Russell, K. C. Heterocycles 1999, 51, 13–16.

Inorganic Chemistry, Vol. 49, No. 24, 2010

11335

1,6-bisphenyl-hexa-1,5-diyne-3,4-dione,62 and bptt63,64 were synthesized using modified literature procedures. Physical Measurements. Samples were prepared using volumetric glassware, and standard Schlenk and drybox techniques. Solvents were degassed by either purging with N2 for >1 h, or several freeze-pump thaw cycles. 1H NMR spectra were collected on a Varian Inova 400 NMR spectrometer with the proton-solvent as a reference. Elemental analyses were performed by Atlantic Microanalytical Laboratory, Norcross, GA. Electronic absorption spectra were collected on a PerkinElmer Lambda 19 UV/Vis/NIR spectrometer at ambient temperature. Luminescence measurements were obtained with a Perkin-Elmer LS 50B luminescence spectrometer equipped with a Hamatatsu model R928 PMT. Quantum yields are given for solutions of matched optical density at the excitation wavelength, and are reported relative to the fluorescence of quinine sulfate in 1 N H2SO4 (φ = 0.55, λex = 345 nm),65 for the high energy ligand fluorescence, and [Ru(bpy)3]Cl2 in air-equilibrated water (φ = 0.028 λex = 465 nm)66 for luminescence of metal complexes. Quantum yields have an experimental error of (5%. Solid-state low temperature emission measurements were collected with the sample submerged in liquid nitrogen using a finger dewar. Solution temperatures above ambient were achieved by immersion in a warm water bath. Solution temperatures below room temperature were achieved by immersion in an ice bath or liquid nitrogen slurries of acetonitrile or acetone and are accurate to within (2 °C. The triplet-sensitized luminescence spectrum of 1 was obtained by exciting a 50 μM solution of 1 with 100 μM xanthone (Aldrich) at λex = 355 nm. Samples containing DNA were prepared via dilution of a stock solution of calf thymus DNA (Sigma) in a buffer of 10 mM Tris-HCl (pH = 7.8) which gave an A260:A280 ratio of 1.9:1, indicating that the DNA was sufficiently free of protein.67 DNA concentrations are expressed as the concentration of bases, which was determined by measurements of the optical density at 260 nm (ε = 6,600 M-1 cm-1 per base).68 For the DNA binding affinity determination, the luminescence spectra of a 50 μM solution of 3 in 10 mM Tris-HCl (pH = 7.8) containing 200 mM NaCl and 10 mM MgCl2 were recorded as a function of increasing CT-DNA concentration (λex = 399 nm) with a Perkin-Elmer LS 50B luminescence spectrometer. A binding isotherm was constructed using the average luminescence intensity at 600 nm for three trials and was fit using the Kaleidagraph software package to a one site binding model on a per base pair basis. The observed luminescence intensity is the sum of the luminescence from the free and bound states and can be defined as I = If þ θIb, where I is the measured luminescence intensity, If, and Ib are the luminescence intensity of the free and bound forms of 3, and θ is the fraction of 3 bound. The fraction bound is equal to R 3 D/Rt, where R 3 D is the concentration of 3 bound to DNA and Rt is the total concentration of 3. In the equilibrium expression for Kd, the dissociation constant is expressed in terms of Rt and Dt, where Dt is the total concentration of DNA bp, since those are known quantities and R 3 D is not directly measured. The value of θ is then obtained directly from these parameters (eq 1). qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðRt þ Dt þ K d Þ - ðRt þDt þK d Þ2 - 4Rt Dt I ¼ If þ Ib ð1Þ 2Dt Because the experimental concentrations are on the order of Kd, no additional assumption could be made to simplify the expression. If, (64) Faust, R.; Ott, S. J. Chem. Soc., Dalton Trans. 2002, 1946–1953. (65) Guilbault, G. G. Practical Fluorescence, 2nd ed.; M. Dekker: New York, 1990; p 812. (66) Nakamaru, K. Bull. Chem. Soc. Jpn. 1982, 55, 2697–2705. (67) Marmur, J. J. Mol. Biol. 1961, 3, 208–218. (68) Reichmann, M. E.; Rice, S. A.; Thomas, C. A.; Dotty, P. J. Am. Chem. Soc. 1954, 76, 3047–3053.

11336 Inorganic Chemistry, Vol. 49, No. 24, 2010 Ib, and Kd were determined from the fit, and the binding association constant obtained by taking the reciprocal of the Kd. Time-resolved luminescence and absorption measurements were performed using the third harmonic of a Coherent Infinity 40-100 pulsed Nd:YAG laser operating at a repetition rate of 10 Hz, (fwhm =5 ns). The laser ouput was used to pump a Lambda Physik Scanmate OPPO (λ = 450 nm) or was Ramanshifted (λ = 396.0 or 447.7 nm) using 500 psi methane gas contained with a cylindrical chamber 1 m in length and 0.06 m in diameter. The laser beam was focused in the center of the chamber, and a Pellin-Broca prism was used to separate and select the desired output wavelength. For transient absorption measurements, the probe beam originated from a PTI model LPS-220 water cooled 150 W Xe arc lamp that was chopped at a rate of 10 Hz (fwhm ∼15 ms). The pump and probe beams converged on a 2 mm quartz cell at a ≈ 10° angle. The probe beam was then focused onto the entrance slit of an ISA model 340S single monochromator equipped with two gratings (1200 grooves/mm and 600 grooves/mm) and two output ports. Kinetics measurements were made utilizing the 1200 groove/mm (500 nm blaze) grating and a Hamamatsu R928P photomultiplier tube. The output from the photomultiplier was fed through a variable gain, broadband amplifier and recorded on a Tektronix (TDS 380) digital oscilloscope. For emission kinetics measurements, a 90° sampling arrangement was utilized where the diffuse emission was focused into the single monochromator. Spectral measurements were performed utilizing the 600 groove/mm (500 nm blaze) grating and a gated ICCD (Princeton Instruments) with a 24.8 mm  6.2 mm active area (1024  256 pixel array). All reported lifetimes have an error of (5%. Solid-state Raman spectra were collected with a Renishaw 1000B micro-Raman spectrometer operating with a λ = 785 nm SDL diode laser. Resonance Raman spectra were collected using Arþ (λ = 457.9, 488.0, and 501.7 nm) and Krþ (λ = 406.7 nm) ion lasers (Coherent models I-70 and I-300, respectively) operating at 100 mW (50 mW at sample). Plasma lines were removed using 600 (λ = 457.9, 488.0, and 501.7 nm) or 1200 grooves/mm (λ = 406.7 nm) gratings prior to the sample. The backscattering was collected with a Nikkor 85 mm 1:1.4 lens focused through a depolarizer (CVI) onto a custom-built f/4 subtractive double monochromator equipped with 600 groove/mm gratings (500 nm blaze). The output from the double monochromator was focused onto the entrance slit of an Action Spectropro 500i monochromator (f/6.5) operating with a 1800 groove/mm (500 nm blaze) for spectra collected using λ = 457.9, 488.0, and 501.7 nm and a 2400 groove/ mm (400 nm blaze) for spectra collected utilizing λ = 406.7 nm excitation. The disperse scattering was then focused onto a backilluminated LN2 cooled CCD (Princeton Instruments) with a 30 mm 14.4 mm active area (2500  600 pixel array). Synthesis of 1,6-Bisphenyl-hexa-1,5-diyne-3,4-dione.62 A 2.85 mL portion of n-butyllithium (1.6 M in hexanes) was added to a stirring solution of 0.50 mL (4.55 mmol) of phenylacetylene in 5 mL of tetrahydrofuran at -78 °C, and the resulting mixture was stirred for 30 min at -78 °C. This solution was added to a mixture of 653 mg (4.55 mmol) of CuBr and 790 mg (9.10 mmol) of LiBr dissolved in 30 mL of tetrahydrofuran at 0 °C. To this reaction mixture, 5.75 mL (2.27 mmol) of oxalyl chloride in 30 mL of tetrahydrofuran was added dropwise. The solution was stirred for 1 h at 0 °C and subsequently quenched with 100 mL of saturated aqueous NH4Cl. The crude product was extracted with ethyl acetate (3  100 mL) and dried over Na2SO4. The solution was then filtered and concentrated under vacuum. The crude product was purified by flash chromatography (4:1 hexanes:ethyl acetate). Yield: 455 mg (78%). 1 H NMR (CD3OD, δ ppm): 7.31 (d, 2H); 7.41 (d, 4H); 7.58 (dd, 4H). 13C NMR (CD3OD, δ ppm): 86.0, 99.9, 118.9, 128.5, 131.6, 133.7, 172.3. Synthesis of 1,6-Bis-(4-tert-butyl-phenyl)-hexa-1,5-diyne3,4-dione. This compound was prepared in a manner identical to 1,6-bisphenyl-hexa-1,5-diyne-3,4-dione. The crude product was

Spencer et al. purified by flash chromatography (10:1 hexanes:ethyl acetate) to give a bright yellow solid in 66% yield. 1H NMR (CDCl3, δ ppm): 7.65 (d, 4H); 7.46 (d, 4H); 1.35 (s, 18H). 13C (CDCl3, δ ppm): 31.0, 35.2, 86.2, 100.8, 116.1, 125.9, 133.9, 155.9, 172.7. HRMS (EI): Calcd for C26H26O2 370.4834. Found 370.4845 (Mþ). Synthesis of 2,3-Bis(phenylethynyl)-1,4,8,9-tetraaza-triphenylene (bptt, 1).63,64 A sample of 0.061 g (0.29 mmol) of 1,6bisphenyl-hexa-1,5-diyne-3,4-dione and 0.075 g (0.29 mmol) 5,6diamino-1,10-phenanthroline were dissolved in 3 mL of glacial acetic acid. The resulting solution was stirred for 5 min and then filtered. The pale yellow solid was washed with 2 mL of acetic acid then 5 mL of methanol and dried under vacuum. Yield: 123 mg (98%). Mp: 292 °C. 1H NMR (CD2Cl2, δ ppm): 7.48 (m, 6H); 7.76 (m, 4H); 7.82 (dd, 2H); 9.24 (dd, 2H); 9.49 (dd, 2H). 13C NMR (CD2Cl2, δ ppm): 86.7, 96.9, 122.1, 124.2, 126.2, 129.1, 130.2, 132.6, 133.6, 138.2, 141.9, 148.1, 152.8. MS (EI): m/z 432.1 (Mþ). Anal. Calcd for C30H16N4: C, 83.32; H, 3.73; N, 12.95. Found: C, 83.21; H, 3.82; N, 12.76. Synthesis of 2,3-Bis(4-tert-butyl-phenylethynyl)-1,4,8,9-tetraaza-triphenylene (tbptt, 2). 5,6-Diamino-1,10-phenanthroline (0.127 g, 0.604 mmol) and 1,6-bis-(4-tert-butyl-phenyl)-hexa-1,5diyne-3,4-dione (0.228 g, 0.616 mmol) were dissolved in 5 mL of glacial acetic acid. The resulting solution was stirred for 5 min and then filtered. The bright yellow solid was washed with acetic acid, then methanol, and dried under vacuum. Yield: 250 mg (76%). 1 H NMR (CDCl3, δ ppm): 9.40 (dd, 2H); 9.20 (d, 2H); 7.77 (dd, 2H); 7.67 (d, 4H); 7.48 (d, 4H); 1.32 (s, 18H). 13C NMR (CDCl3, δ ppm): 31.2, 35.3, 86.7, 97.3, 118.8, 124.4, 126.2, 126.7, 132.4, 133.6, 138.3, 141.4, 147.9, 152.6, 154.0. HRMS (EI): Calcd for C38H32N4 544.26270. Found 544.26273 (Mþ). Synthesis of [Ru(phen)2bptt](PF6)2 (3). To a stirring solution of 0.185 g (0.347 mmol) of Ru(phen)2Cl2 in 20 mL of methanol, 0.150 g (0.347 mmol) of 1 and 0.131 g (0.764 mmol) of NaCF3SO3 were added, and the reaction mixture was allowed to reflux for 2 h. The NaCl precipitate was filtered, and 25 mL of saturated aqueous NH4PF6 was added to the filtrate to precipitate the product. The solution was filtered, and the solid was dried under vacuum and recrystallized from methanol. Yield: 150 mg (37%). X-ray quality crystals were grown from a saturated CH2Cl2 solution layered with ethanol. 1H NMR (DMSO, δ ppm): 9.38 (d, 2H); 8.79 (d, 2H); 8.77 (d, 2H); 8.40 (s, 4H); 8.23 (dd, 4H); 8.06 (d, 2H); 7.86-7.73 (m, 10H); 7.58-7.52 (m, 6H). 13C NMR (DMSO, δ ppm): 86.4, 97.7, 120.1, 126.3, 126.4, 127.5, 128.1, 128.5, 129.3, 130.5, 130.9, 132.1, 133.2, 137.0, 137.6, 141.1, 147.1, 149.7, 152.7, 153.3, 154.4. MS (ESI): m/z 1039.4 [M - PF6]þ. Anal. Calcd for C54H32N8P2F12Ru: C, 54.78; H, 2.72; N, 9.46. Found: C, 54.52; H, 2.82; N, 9.22. Synthesis of [Ru(phen)2tbptt](PF6)2 (4). Complex 4 was prepared utilizing a modified literature procedure.64 To a degassed mixture of 20 mL of ethanol, 5 mL of dichloromethane, and 0.115 g (0.211 mmol) of 2, 113 mg (0.200 mmol) of Ru(phen)2Cl2 was added. The resulting solution was heated to 100 °C and allowed to stir for 18 h under nitrogen. The mixture was then cooled to 0 °C, and 3 mL of a saturated aqueous solution of NH4PF6 was added to generate a red-orange precipitate. The solution was filtered, and the precipitate washed with water, ethanol, and ether, then recrystallized from methanol to give 4. Yield: 205 mg (79%). 1H NMR (DMSO, δ ppm): 9.39 (d, 2H); 8.79 (d, 2H); 8.77 (d, 2H); 8.40 (s, 4H); 8.25 (d, 2H); 8.21 (d, 2H); 8.06 (d, 2H); 7.82 (m, 6H); 7.68 (d, 4H); 7.55 (d, 4H); 1.31 (s, 18H). 13C NMR (DMSO, δ ppm): 30.8, 34.9, 86.1, 98.0, 117.3, 126.1, 127.4, 128.1, 128.5, 130.5, 132.0, 137.0, 137.3, 141.2, 147.1, 149.6, 152.7, 153.4, 153.8, 154.3. MS (ESI): m/z 1151.1 [M - PF6]þ. Anal. Calcd for C62H48F12N8P2Ru: C, 57.45; H, 3.73; N, 8.65; Found: C, 57.52; H, 3.72; N, 8.52. Crystallographic Structure Determination of 3. An orange crystal of 3 (approximate dimensions 0.08  0.005  0.003 mm3) was placed onto the tip of a 0.001 mm diameter glass capillary and mounted on a Kappa SMART setup equipped with a 2  2 array of 1K CCD detectors at beamline 15ID, ChemMatCARS, APS

Article Chicago, for a data collection at 123(2) K. The data collection was carried out using λ = 0.56356 A˚ (energy: 22.113 keV, monochromator: two diamond 1 1 1 crystals, using two mirrors to exclude higher harmonics) with a frame time of 5 s and a detector distance of 6 cm. Frames were collected with 0.30° steps in φ and at 0° for all other angles. The data reduction was carried out with SAINT69 using the orientation matrix of the major twin component and including data to a resolution of 0.8238 A˚ (40° for silver radiation). The space group P1 was determined based on intensity statistics and the lack of systematic absences. A direct-methods solution was calculated which provided most non-hydrogen atoms from the E-map.70 Full-matrix least-squares/difference Fourier cycles were performed which located the remaining non-hydrogen atoms.71 All non-hydrogen atoms of the Ru-complex were refined with anisotropic displacement parameters. The data were corrected for nonmerohedral twinning72 (twin law by rows -1 0 0, 0.291 1 -1.028, 0 0 -1) in the final refinements, which resulted in significantly decreased R-values and a clean difference map. The structure was found with two independent molecules in the asymmetric unit, related by pseudo-symmetry. The two independent molecules were restrained to be geometrically the same within 0.001 A˚. The packing plot shows that the ruthenium complexes π-stack along the a-axis with distances between the pyrazine rings of 3.51 A˚ between the independent molecules and 3.83 and 3.75 A˚ between molecule A and its symmetry equivalent and molecule B and its symmetry equivalent, respectively. Computational Methods and Vibrational Analysis. Geometry optimizations for 1 and 3 were performed using density functional theory (DFT) and the B3LYP combination of functionals available in Gaussian 98.73-75 A 6-31G* basis set was used for all calculations involving 1. For 3, the LANLDZ effective core potential was used for Ru2þ and a 3-21G basis set was used for the C, N, and H atoms.76-78 Since the crystal structure of 3 was not an energetic minimum in the gas phase, only the vibrational analysis at the gas phase B3LYP optimized geometry of 3 is presented here. The DFT vibrational frequencies were scaled by 0.9497 resulting in values that were in good agreement with the experimental results for known vibrations such as -CtC-. For 1, time-dependent DFT (TDDFT) calculations were performed with the same functionals and basis set at the optimized geometry to obtain electronic transition energies.

Results and Discussion Synthesis. The dione precursors to 1 and 2 were obtained utilizing a modified literature procedure.62 Slow addition of the appropriate lithium-phenylacetylene salt to a CuBr (69) Glazer, E. C.; Magde, D.; Tor, Y. J. Am. Chem. Soc. 2007, 129, 8544– 8551. (70) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.; Burla, M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27, 435. (71) Sheldrick, G. Acta Crystallogr., Sect. A 2008, 64, 112–122. (72) Young, V. ROTWIN, Program for data correction for non-merohedral twins; unpublished. (73) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A. J.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Laishenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Callacombe, M.; Gill, P. M. W.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.6; Gaussian,Inc.: Pittsburgh, PA, 1998. (74) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter 1988, 37, 785–789. (75) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (76) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283. (77) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310. (78) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284–298.

Inorganic Chemistry, Vol. 49, No. 24, 2010

11337

Figure 1. ORTEP of the X-ray crystal structure of 3. Thermal ellipsoids are illustrated at 30% probability.

solution followed by the addition of oxalyl chloride results in yields of ∼65-80% of the dione precursors. The subsequent condensation reaction of the dione with 5,6-diamino1,10-phenanthroline63,64 afforded ligands 1 and 2 in 98% and 76% yield, respectively. The complexation reactions were performed via addition of the ligand to a solution of Ru(phen)2Cl2. The higher yield of 4 primarily derives from the use of solvent conditions suitable for complete dissolution of 2 (4:1 EtOH:CH2Cl2) and longer reaction times (18 h). X-ray Crystal Structure of 3. Crystals suitable for X-ray structure determination were obtained by layering and slow diffusion of ethanol into a saturated dichloromethane solution of 3. Crystals of 3 were small and required evaluation using the synchrotron source at Argonne National Laboratory. The crystal was twinned and poorly diffracting leading to a lower quality structure, but it was sufficient for the purposes of the present discussion. The crystallographic data for 3 are presented in the Supporting Information, Table S1. An ORTEP illustration of the crystal structure is shown in Figure 1. The Ru(II)-center possesses a distorted octahedral coordination sphere as a consequence of the small bite angles (average = 79.2°) of the bidentate ligands. The mean Ru-N bond length (2.08 A˚) falls within the range typical for [Ru(diimine)3]2þ complexes.79-86 It is slightly longer than those reported for [Ru(phen)3](PF6)2 (2.07 A˚)81 and [Ru(bpy)3](PF6)2 (2.056 A˚)79 but shorter than those reported for [Ru(Me2phen)2dppz](PF6)2 (2.10 A˚).83 The C31-C47C48 (175.4° (13)) and C30-C39-C40 (173.9° (11)) bond angles of the enediyne unit do not significantly deviate from linearity. The distance between the alkyne termini (4.17 A˚) is large, but typical for stable acyclic enediyne motifs.87,88 The (79) Rillema, D. P.; Jones, D. S.; Levy, H. A. Chem. Commun. 1979, 849–851. (80) Rutherford, T. J.; Pellegrini, P. A.; Aldrich-Wright, J.; Junk, P. C.; Keene, F. R. Eur. J. Inorg. Chem. 1998, 1677–1688. (81) Maloney, D. J.; MacDonnell, F. M. Acta Crystallogr., Sect. C 1997, C53, 705–707. (82) Komatsuzaki, N.; Katoh, R.; Himeda, Y.; Sugihara, H.; Arakawa, H.; Kasuga, K. J. Chem. Soc., Dalton Trans. 2000, 3053–3054. (83) Liu, J.-G.; Zhang, Q.-L.; Shi, X.-F.; Ji, L.-N. Inorg. Chem. 2001, 40, 5045–5050. (84) Greguric, A.; Greguric, I. D.; Hambley, T. W.; Aldrich-Wright, J. R.; Collins, J. G. J. Chem. Soc., Dalton Trans. 2002, 849–855. (85) Berg-Brennan, C.; Subramanian, P.; Absi, M.; Stern, C.; Hupp, J. T. Inorg. Chem. 1996, 35, 3719–3722. (86) Breu, J.; Stoll, A. J. Acta Crystallogr., Sect. C 1996, C52, 1174–1177. (87) Rawat, D. S.; Benites, P. J.; Incarvito, C. D.; Rheingold, A. L.; Zaleski, J. M. Inorg. Chem. 2001, 40, 1846–1857. (88) Kraft, B. J.; Coalter, N. L.; Nath, M.; Clark, A. E.; Siedle, A. R.; Huffman, J. C.; Zaleski, J. M. Inorg. Chem. 2003, 42, 1663–1672.

11338 Inorganic Chemistry, Vol. 49, No. 24, 2010 ancillary phenanthroline ligands are planar to within 0.07 A˚, with a dihedral angle of 96° between the two planes. The coordinated bptt ligand is slightly distorted, possessing a bowed structure where the apex corresponds to the pyrazine ring. This distortion is likely a product of the π-stacking interaction observed at the pyrazine ring (3.51 A˚ between independent molecules) because it is not observed in the gas phase structure optimized at the 3-21G level. The effect of the π-stacking interaction on the bonding within the pyrazine ring is reflected in an increase of the N7-C29 and N8-C32 bond lengths and a corresponding decrease in the C29-C32 and C30-C31 bond lengths relative to their computed values (see Supporting Information, Table S2). A slight decrease in the in the C-C bond order facilitates out-ofplane distortions at the pyrazine ring, leading to the bowing of the bptt ligand observed in the X-ray structure. Electronic Structure. Both bptt derivatives show intense ππ* transitions at λ ∼ 395 nm and a weak nπ* feature centered at λ ∼ 460 nm (Figure 2a). The λ = 393 nm peak (ε393 = 33,000 M-1 cm-1) in 1 is red-shifted to λ = 398 nm in 2 by incorporation of a tert-butyl group at the para position of the terminal phenyl ring (ε398 = 33,200 M-1 cm-1), while the weaker features centered at λ ∼ 460 nm do not mirror this effect. Free ligands 1 and 2 show notably strong fluorescence at room temperature, with λmax = 423 and 439 nm, respectively. The shift loosely correlates to the changes in the absorption spectrum of the ligand upon substitution. A parallel red-shift is also observed in the 77 K phosphorescence spectra of 1 and 2 where the maxima of λ = 531 and 568 nm for 1 are shifted to λ = 535 and 575 nm for 2. These observations indicate that the frontier molecular orbitals of 1 and 2 have significant character on the terminal phenyl groups, while the weaker nπ* features are predominantly localized on the pyrazine ring. Support for this interpretation derives from time-dependent density functional theory (TDDFT) calculations (Figure 3). The lowest energy allowed electronic transitions calculated for 1 involve π and π* orbitals primarily localized on the pyrazine, dialkynyl, and terminal phenyl groups. Excitation of these transitions leads to antibonding character at the alkynes and within the pyrazine ring. From the density plots, it is clear that there is a much larger contribution from the para position of the terminal phenyl groups in the HOMO and SHOMO than in the LUMO and SLUMO. Substitution of a strong σ-donor group at this position should therefore destabilize the HOMO and have a similar, but smaller, effect on the LUMO energy, resulting in a small red shift of the electronic transition. A red shift of the λ ∼ 395 nm transition is indeed observed upon substitution (Figure 2a) indicating that the TDDFT result is consistent with the electronic spectra of 1 and 2 in solution. The electronic absorption spectra of 3 and 4 have two prominent low energy features; a ligand-centered transition at λ ∼ 395 nm and metal-to-ligand charge transfer transition at λ ∼ 450 nm (Figure 2b). The red-shift in the λ ∼ 395 nm feature observed in the electronic spectra of 1 and 2 upon tert-butyl substitution (Figure 2a) is retained between complexes 3 and 4 (Figure 2b). These ligandcentered transitions also appear to gain intensity (3: ε387 = 41,500 M-1 cm-1, 4: ε394 = 42,100 M-1 cm-1), but this is a result of overlap with the Ru(dπ)-to-phen(π*) charge-transfer transitions and the underlying Ru2þ d-d bands that also

Spencer et al.

Figure 2. Electronic absorption and emission spectra for 1-4 collected in ethanol. (a) 298 K absorption and fluorescence (λex = 395 nm, λem ∼ 420 nm) spectra and 77 K phosphorescence (λex = 395 nm, λem ∼ 530 nm) for 1 (black) and 2 (red). (b) 298 K absorption and luminescence (λex = 450 nm, λem ∼ 600 nm) spectra for 3 (black) and 4 (red). Fluorescence (φ1 and φ2) and luminescence (φ3 and φ4) quantum yields for 1-4 are shown above the emission profiles.

Figure 3. TDDFT description of the frontier molecular orbitals of 1.

occur in this energy region.89 Surprisingly, the primary charge transfer transition at λ ∼ 450 nm remains largely unaffected by the substitution at the terminal phenyl rings. Given the effect of tert-butyl substitution on the lowest energy π-π* transition in 1 and 2, one would expect a (89) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85–277.

Article

Inorganic Chemistry, Vol. 49, No. 24, 2010

11339

Table 1. Vibrational Assignments from Density Functional Theory and Experimental Resonance Raman Shifts for 3 ν

assignmenta

B3LYP/3-21G// LANL2DZ (cm-1)b

B3LYP/3-21G// LANL2DZ (cm-1)c

80 89 96 100 112 122 166 175 187 189 198 199 203 206 222 226 238 240 241 242 246

alkyne oop-wag phenyl breathe bptt-phen breathe ancillary-phen breathe ancillary-phen H-oop wag ancillary-phen breathe pyrazine breathe ancillary-phen breathe bptt-phen C-C stretch ancillary-phen C-N stretch ancillary-phen C-C stretch bptt-phen C-C stretch ancillary-phen C-C stretch pyrazine C-N stretch bptt-phen C-C stretch ancillary-phen C-C stretch bptt-phen C-C stretch ancillary-phen C-C stretch ancillary-phen C-C stretch phenyl C-C stretch alkyne stretch

620 684 749 761 840 933 1128 1227 1260 1287 1327 1339 1378 1417 1515 1536 1599 1609 1610 1626 2321

588 649 711 723 798 886 1071 1165 1197 1222 1260 1272 1309 1346 1438 1458 1519 1528 1529 1544 2204

a

Raman shift (cm-1) λex = 457. Nine nm

Raman shift (cm-1) λex = 406.7 nm

727 805 883 1166

590 654 708 725 1070 1165 1211 1221

1269 1272 1319 1349 1435 1457 1568 1572 1599 2204

1595 2204

Assignments represent the region of the molecule with the greatest atomic displacement. b Unscaled values. c Scaled by 0.9496.

corresponding shift in the MLCT transitions of 3 and 4 if the ligand LUMO is delocalized about the bptt π-system. Additionally, given the low reduction potential reported for 3 (-1.14 V vs Fc/Fcþ),64 a significantly red-shifted MLCT transition relative to the [Ru(bpy)3]2þ and [Ru(phen)3]2þ derivatives would be expected.89,90 The absence of a shift in the MLCT transition upon tert-butyl substitution suggests that 3 and 4 have bichromophoric character, similar to [Ru(phen)2dppz]2þ and derivatives thereof.47,52,91-94 This is most clearly depicted by the DFT description of the frontier orbitals of 1 (Figure 3), which indicates that the bptt ligand is composed of two non-degenerate π-systems where the LUMO and SLUMO are primarily localized on the pyrazine, dialkynyl, and terminal phenyl groups, while the phenanthroline π*-orbitals lie higher in energy. Although it might be expected that the phen-based π*-orbitals of 1 and 2 would be sufficiently stabilized by complexation to facilitate communication between these two π-systems,47 this effect is not apparent in the ground-state electronic absorption spectra of 3 and 4. The bichromophoric description of the electronic structure derived from the electronic spectrum of 3 (Figure 2b) is consistent with the ground-state resonance Raman spectra, and corresponding DFT vibrational assignments (Table 1). Generally, vibrational modes associated with the bptt ligand are enhanced upon λex = 406.7 nm, while vibrations associated with the ancillary phenanthrolines are enhanced with λex = 457.9 nm (Table 1). With λex = 457.9 nm excitation, the modes at 1457, 1572, and 1599 cm-1 associated with the ancillary phenanthrolines all show significant enhancement, while bptt modes such as the (90) Balzani, V.; Credi, A.; Venturi, M. Coord. Chem. Rev. 1998, 171, 3–16. (91) Ackermann, M. N.; Interrante, L. V. Inorg. Chem. 1984, 23, 3904– 3911. (92) Amouyal, E.; Homsi, A.; Chambron, J. C.; Sauvage, J. P. J. Chem. Soc., Dalton Trans. 1990, 1841–1845. (93) Chambron, J. C.; Sauvage, J. P.; Amouyal, E.; Koffi, P. Nouv. J. Chim. 1985, 9, 527–529. (94) Fees, J.; Ketterle, M.; Klein, A.; Fiedler, J.; Kaim, W. J. Chem. Soc., Dalton Trans. 1999, 2595–2600.

alkyne stretch at 2204 cm-1 show very low intensity (Figure 4b). Conversely, with λex = 406.7 nm, vibrations associated with the bptt ligand (ν89, ν189, ν246) dominate the vibrational spectrum (Figure 4a). This includes modes associated with the phenanthroline (ν96, ν199, ν222, and ν238), the pyrazine-dialkynyl (ν80, ν166, ν206, and ν246), and the terminal phenyl (ν89 and ν242) fragments of the bptt ligand. Enhancement of vibrations distributed throughout the bptt ligand upon λex = 406.7 nm excitation implies an electronic structure of the coordinated ligand that is perturbed significantly from that described in Figure 3. Specifically, the Raman spectrum suggests that the electronic transition at λ ∼ 395 nm in 3 is composed of a more extensively delocalized orbital involving both the phenanthroline and the pyrazine-dialkynyl portions of the ligand. A more delocalized orbital may derive from stabilization of the phenanthroline-centered π* orbitals of bptt upon complexation which facilitates a mixing of the non-degenerate phenanthroline and pyrazine-dialkynyl orbitals. Such delocalization would result in a redshift of the MLCT transition relative to the electronic spectrum of [Ru(bpy)3]2þ. This is clearly not observed in the ground-state electronic absorption spectrum of 3 (Figure 2b). The enhancement pattern in the λ = 406.7 nm Raman spectrum most likely derives from a degeneracy of two distinct electronic transitions (ππ* bptt-pyrazine/ dialkynyl-centered and Ru(dπ)-bptt(π*)) which results in detection of vibrations from both fragments of the bptt ligand. To a first approximation, the room temperature emission spectra of the ruthenium complexes 3 and 4 demonstrate behavior comparable to most ruthenium polypyridine derivatives upon MLCT excitation (Figure 2b).89 The emission maxima (λ=603 nm for 3 and 4) and quantum yields of 3 (φ = 0.066) and 4 (φ = 0.056) are relatively unaffected by substitution at the terminal phenyl of the bptt ligand and are similar to those reported for [Ru(phen)3]2þ (λ = 604 nm, φ = 0.058).66 In contrast, the room temperature lifetimes (τ3 = 630 ns, τ4 = 590 ns) are significantly

11340 Inorganic Chemistry, Vol. 49, No. 24, 2010

Figure 4. Resonance Raman spectra of 3 collected at 298 K in DMSO (20 mM) with (a) λex = 406.7 nm, and (b) λex = 457.9 nm.

shorter than those reported for [Ru(phen)3]2þ (τ ∼ 1.0 μs).89 This discrepancy implies quenching by the bptt ligand, similar to what is observed for [Ru(phen)2dppz]2þ in polar or H-bonding solvents.42,89 Both 3 and 4 are structurally and electronically related to [Ru(phen)2dppz]2þ, as they possess two non-degenerate low-energy electronic transitions. The transient absorption spectra of 3 and 4 also imply an influence of the bptt ligand on the nature of the lowest-lying excited state (Figure 5). The transient absorption spectral features for 3 and 4 decay with the same lifetime as the luminescence decay (τ3 = 650 ns and τ4 = 600 ns) and are generally similar to those observed for [Ru(bpy)3]2þ.92,95 Photoexcitation (λex=450 nm) of 3 leads to bleaching of both the MLCT and the ligand-centered transitions at λ ∼ 450 nm and λ ∼ 390 nm, and an increase in the OD at λ ∼ 345 nm. For 4, the λ ∼ 345 nm peak and the λ ∼ 390 nm bleach show a ∼ 5 nm red-shift upon tert-butyl substitution, similar to the trend observed in the ground state absorption spectra of 1-4. The primary difference between the literature transient absorption spectra for [Ru(bpy)3]2þ and spectra for 3 and 4 is the position of the positive feature at λ ∼ 345 nm, which is significantly blue-shifted relative to [Ru(bpy)3]2þ (λ ∼ 375 nm). This shift is likely related to the bleach at λ ∼ 390 nm which influences the observed maximum of the positive feature at λ ∼ 375 nm in [Ru(bpy)3]2þ. Given these spectral differences and the quenched lifetimes of 3 and 4 relative to [Ru(bpy)3]2þ, it is apparent that the emissive state (or states)96,97 has character that extends out to the terminal phenyl groups of the bptt ligand. Interactions with DNA. Given the structural and electronic similarities of 3 and 4 to widely studied [Ru(phen)2dppz]2þ, we have focused our investigations on the electronic effects of DNA binding on the emission energies and lifetimes of 3 and 4 in an aqueous environment. Similar to the electronic absorption spectrum in ethanol, the spectrum of 3 in an aqueous environment contains two low-energy transitions at λ ∼ 450 nm and (95) Bensasson, R.; Salet, C.; Balzani, V. J. Am. Chem. Soc. 1976, 98, 3722–3724. (96) Hager, G. D.; Crosby, G. A. J. Am. Chem. Soc. 1975, 97, 7031–7037. (97) Sykora, M.; Kincaid, J. R. Inorg. Chem. 1995, 34, 5852–5856.

Spencer et al.

Figure 5. Transient absorption spectra of 3 (black) and 4 (red) collected between 50 and 150 ns after laser excitation (λex = 450 nm) in ethanol at 298 K. Inset: Change in optical density of 3 at 350 nm after λex = 450 nm.

λ ∼ 390 nm (Figure 6a). The primary spectral difference between the two conditions (i.e., in ethanol and in water) is the marked decrease in the intensity of the 390 nm transition (εmax = 29,100 M-1 cm-1), as well as a slight red-shift of the peak maximum (λmax = 395 nm). Similar changes are observed for the tert-butyl derivative, 4, although the decrease in intensity and red-shift of the band maximum results in significant spectral overlap between the ligand-centered and MLCT transitions, and thus poor resolution of these spectral features (Figure 6b). In addition to the changes in the absorption spectrum, the emission spectra of both 3 and 4 are significantly red-shifted (3: λem = 634 nm; 4: λem = 634 nm) and dramatically quenched, similar to what is observed for [Ru(phen)2dppz]2þ.42,45 Addition of 10 mol equiv (nucleobase) of CT-DNA to a solution of 3 induces a decrease in the intensity and a red shift of the 395 nm transition (λmax = 400 nm, εmax = 26,400 M-1 cm-1), concurrent with a 9-fold increase in the luminescence quantum yield (Figure 6a). These effects are similar to the spectral signatures of intercalation for [Ru(phen)2dppz]2þ and related complexes.9,33,83,98 In contrast to 3, 4 demonstrates only minor changes in both the electronic absorption and the emission spectra upon addition of CT-DNA (Figure 6b), suggesting tert-butyl substitution at the terminal phenyl dramatically decreases the ability of 4 to intercalate between the DNA base pairs. Additionally, the relatively modest increase in luminescence quantum yield for 3 relative to dppz (>104)33 reveals that the intercalation of 3 occurs via π-stacking involving the terminal phenyl groups, leaving the pyrazine portion of the molecule partially accessible for solvent quenching. The intrinsic luminescence of 3 can be monitored to quantitatively evaluate its binding affinity for DNA in aqueous buffer relative to [Ru(phen)2dppz]2þ. Relatively high salt concentrations (200 mM NaCl and 10 mM MgCl2) were employed to discourage non-specific electrostatic binding. The emission maximum (λem = 600 nm) is red-shifted relative to that of 3 in Tris-HCl buffer with (98) Hartshorn, R. M.; Barton, J. K. J. Am. Chem. Soc. 1992, 114, 5919– 5925.

Article

Inorganic Chemistry, Vol. 49, No. 24, 2010

11341

Figure 7. Binding isotherm for 3 binding to CT-DNA determined by intrinsic luminescence at 600 nm (λex = 399 nm). Inset: Representative luminescence emission spectra of 3 as a function of increasing CT-DNA.

Figure 6. Electronic absorption and emission (λex = 450 nm) spectra of 3 (a) and 4 (b) in the absence (black) and presence (red) of 10 mol equiv (nucleobase) CT-DNA in 10 mM Tris-HCl at pH 7.2 and 298 K. Luminescence quantum yields for 3 and 4 in the absence (φH20) and presence (φDNA) of DNA are shown above the emission profiles.

no added salts (λem = 615 nm). The emission intensity at λ = 600 nm after excitation at λ = 399 nm was used to construct a DNA binding curve (Figure 7). The luminescence intensity increases by a factor of nearly 10 over the course of the experiment, consistent with the previous increase observed for the 10-fold excess CT-DNA measurement. This is accompanied by a small shift of the emission maxima (∼5 nm) to longer wavelength. Fitting the binding isotherm to a simple one site binding model on a per base pair basis with no additional assumptions yields a binding association constant (Kb) of 3.3 ( 0.2  104 M-1. The strength of this interaction in relatively high salt concentration indicates the binding is intercalative. This Kb is in good agreement with those reported for other molecular light switch complexes which range from 104 to 107 M-1.28,33,36,99 Specifically, the Kb was determined to be 4.95  106 M-1 for [Ru(phen)2dppz]2þ and 8.4  105 M-1 for [Ru(bpy)2taptp]2þ by following spectral changes in the electronic absorption profile upon DNA binding.36 Since the bptt ligand is bulkier and less planar than either of these ligands, one would expect to observe a lower affinity intercalative interaction. The luminescence lifetimes of 3 and 4 in an aqueous environment both demonstrate biphasic kinetics (3: τ1 = 40 ns, τ2 = 230 ns; 4: τ1 ∼ 26 ns, τ2 = 150 ns) with significantly reduced lifetimes relative to [Ru(phen)3]2þ (Figures 8a and 8c). Upon addition of DNA, the luminescence lifetime of 3 increases considerably (τ1 = 220 ns, τ2 = 990 ns), but retains biphasic decay kinetics (Figure 8b). (99) Pyle, A. M.; Rehmann, J. P.; Meshoyrer, R.; Kumar, C. V.; Turro, N. J.; Barton, J. K. J. Am. Chem. Soc. 1989, 111, 3051–3058.

The decay kinetics of the transient absorption at λ = 350 nm of 3 in the presence of DNA are also biexponential (τ1 = 210 ns, τ2 = 1000 ns) and possess rate constants that closely parallel those observed for the excited state decay by emission (Figure 8b). In contrast to 3, the lifetime of 4 increases only slightly upon addition of DNA (τ1 = ∼29 ns, τ2 = 260 ns) (Figure 8d), a result that is consistent with the small change observed in the luminescence quantum yield under the two solution conditions. The excited state model of [Ru(phen)2dppz]2þ luminescence decay describes two distinct 3MLCT excited states; a 3 MLCTphen state, where the negative charge is delocalized about phenanthroline components of the three ligands, and a lower-energy 3MLCTphzn state, where the negative charge resides primarily on the phenazine portion of the dppz ligand.31,39,41,45,46 Ultrafast time-resolved emission and resonance Raman experiments have demonstrated that there is rapid interconversion between these two states, and they likely establish an equilibrium condition in less than 20 ps.41,45,49 The energy of the 3MLCTphzn state is highly sensitive to solvent interactions at the pyrazine nitrogens. In polar or H-bonding solvents, the 3MLCTphzn state is stabilized sufficiently to significantly deactivate the 3MLCTphen state leading to a dramatically reduced luminescence lifetime relative to [Ru(phen)3]2þ. In water, [Ru(phen)2dppz]2þ exhibits emission from the 3MLCTphen state at λ ∼ 630 nm and weak emission from the 3MLCTphzn state in the near-IR (λ ∼ 800 nm).45 If the phenazine-based MLCT state of 3 (3MLCTbptt) were emissive, it would be expected to appear at lower energy than the analogous state in [Ru(phen)2dppz]2þ because of the extended conjugation and to have an even lower quantum yield since its reduced rigidity leads to increased non-radiative decay. However, for 3 and 4, only a single emission band corresponding to the 3MLCTphen excited state is observed and no emission features are present between 700 and 1200 nm. Even though the 3MLCTbptt state is non-emissive and does not contribute to the observed dual emission, it is responsible for the light switch effect for 3 upon DNA binding. Though [Ru(phen)2dppz]2þ displays biphasic decay kinetics when bound to DNA, the decay is monoexponential in water. From studies on the luminescence lifetimes of

11342 Inorganic Chemistry, Vol. 49, No. 24, 2010

Spencer et al. Table 2. Luminescence Lifetimes of 3 as a Function of Solvent % acetonitrile in water

τ1/ns

τ2/ns

F1 a

R2b

100 80 60 40 20 0

1040 230 72 38 ∼19 40

n/a n/a n/a n/a 160 230

1 1 1 1 0.23 0.19

0.9998 0.9994 0.9994 0.9993 0.993 0.99

a Fractional contribution to luminescence from τ1. b Square of the correlation coefficient for the data fit.

Figure 8. Kinetic profiles of the excited state decay of 3 and 4 in the presence and absence of 10 mol equiv (nucleobase) CT-DNA in 10 mM Tris-HCl at pH 7.2 and 298 K after λex = 450 nm. (a) Luminescence decay of 3 at λ = 625 nm in the absence of DNA. (b) Transient absorbance (λ = 350 nm) and luminescence decay (λ = 625 nm) of 3 in the presence of DNA. (c) Luminescence decay of 4 at λ = 625 nm in the absence of DNA. (d) Luminescence decay of 4 at λ = 625 nm in the presence of DNA.

enantiomerically pure samples of [Ru(phen)2dppz]2þ, it has been shown that each enantiomer (Λ and Δ) possesses unique biphasic decay kinetics in the presence of DNA (Λτw = 40 ns, Λτt = 150 ns and Δτw = 160 ns, Δτt = 850 ns)100 deriving from differing affinities of the two enantiomers for the handed DNA substrate.33,100 The biexponential emission decays for each enantiomer (Λ or Δ) have been determined to derive from the relative proximity of the intercalated [Ru(phen)2dppz]2þ units.33 This analysis indicates that the longer lifetime component (τt) derives from more closely bound emissive intercalators, which generates a tightly packed DNA-metal complex structure resulting in more efficient solvent shielding of the pyrazine nitrogens.33 In contrast, emissive intercalators that are widely spaced along the DNA polymer experience a less efficient shielding of the pyrazine nitrogens, and consequently demonstrate a shorter lifetime (τw). However, the photophysical behavior of 3 cannot be explained by the same phenomena even though a racemic mixture was employed for binding studies, since it also displays biphasic kinetics in aqueous solution in the absence of DNA where the stereochemistry of the complex would have no effect. Solvent-Dependent Luminescence. In an effort to further explore this unusual biphasic behavior, the solvent dependence of the emission kinetics of 3 was explored in varying CH3CN:H2O solvent mixtures. Samples were excited at λ = 447.7 nm, and the emission kinetics were evaluated at 600 nm. As expected, in pure CH3CN, the lifetime (τ = 1040 ns) is longer than that measured in EtOH, and nearly identical to that of [Ru(phen)3]2þ in CH3CN (Table 2).66 In non-hydrogen bonding solvents, there is little quenching because of interactions at the phenazine nitrogens leading the complex to behave like [Ru(phen)3]2þ. As CH3CN is replaced with H2O in the solvent mixture, increased H-bonding at the phenazine nitrogens causes the lifetimes to decrease (100) Erkkila, K. E.; Odom, D. T.; Barton, J. K. Chem. Rev. 1999, 99, 2777–2795.

(Table 2). When the amount of CH3CN is reduced to 20% of the solution volume, 3 displays biphasic emission kinetics where the fraction of molecules emitting from the state corresponding to the shorter lifetime (F1) is 0.23 which is comparable to that fraction in pure H2O (F1 = 0.19). The somewhat reduced lifetimes in 20% CH3CN versus pure H2O may be due the decreased ability to form H-bonding networks that can stabilize the excited state(s) in the solvent mixture. Overwhelmingly, molecules obey Kasha’s rule and only appreciably emit from the lowest energy excited state.101 However, at low temperature, many molecules exhibit biphasic decay profiles since excited states that would be equilibrated at room temperature can be trapped by reducing the temperature. Many of these systems exhibit simultaneous emission from 3MLCT and ligand-based triplet states and are discussed in a comprehensive review by Schmehl et al.50 Molecules that exhibit biphasic decay kinetics in solution at room temperature have only rarely been reported.69,102-104 In each case, biphasic luminescence is the result of two emitting excited states with similar energies separated by a kinetic barrier that prevents complete equilibration of the two states resulting in dual emission. Two of these reports involve mixed-ligand Ru(II) diimine complexes. In the first case, Zhang and co-workers synthesized a complex, [Ru(bpy)2L]2þ (where L is 4-methyl-40 [p-(dimethyl-amino)-R-styryl]-2,20 -bipyridine) that displayed dual emission at room temperature.103 Using time-resolved emission spectra, they were able to resolve two luminescence maxima at different energies and assign the two emitting states as 3MLCT and 3ILCT in nature. In the second example, Tor et al. synthesized a series of alkyne-substituted phen ligands and corresponding mixed-ligand Ru(II) complexes and found those asymmetrically substituted at the 4-position of the phenanthroline have biphasic emission profiles with the fractional contributions from the shortand long-lived components variable with the detection wavelength.69 The proposed explanation suggests that the 3 MLCT state involving the alkyne-substituted phen ligand was not completely equilibrated with the phen-based 3 MLCT resulting in emission from both 3MLCT states. Attempts to isolate the two emitting states in 3 by examining time-resolved emission spectra were largely unsuccessful. A small red-shift in the emission maxima from 600 to 610 nm is observed at long time delays (>1 μs (101) Kasha, M. Faraday Discuss. 1950, 9, 14–19. (102) Ferraudi, G.; Feliz, M.; Wolcan, E.; Hsu, I.; Moya, S. A.; Guerrero, J. J. Phys. Chem. 1995, 99, 4929–4934. (103) Song, L.; Feng, J.; Wang, X.; Yu, J.; Hou, Y.; Xie, P.; Zhang, B.; Xiang, J.; Ai, X.; Zhang, J. Inorg. Chem. 2003, 42, 3393–3395. (104) Matsumoto, K.; Matsumoto, N.; Ishii, A.; Tsukuda, T.; Hasegawa, M.; Tsubomura, T. Dalton Trans. 2009, 6795–6801.

Article

Inorganic Chemistry, Vol. 49, No. 24, 2010

11343

Table 3. Temperature Dependence of the Luminescence Lifetimes of 3 solvent EtOH

60:40 Ethylene Glycol:H2O

temperature/K

τ1/ns

τ2/ns

F1a

R2 b

298 233 193 77 333

630 1510 1770 4210 143

n/a n/a n/a n/a ∼1190

1 1 1 1 >0.99

0.9998 0.9996 0.9999 0.991 0.999

298 273 233 77

82 61 41 4980

800 1120 1410 n/a

0.93 0.91 0.89 1

0.9992 0.998 0.999 0.9997

a Fractional contribution to luminescence from τ1. b Square of the correlation coefficient for the data fit.

after the laser pulse) for 3 in the presence of DNA in water (λex = 447.7 nm) indicating that the two states are nearly degenerate (data not shown). Furthermore, determination of the decay kinetics across the emission spectrum under the same conditions reveals that the luminescence is biphasic at all emission wavelengths, and the lifetimes are independent of wavelength. However, the fractional component F1 varies across the emission spectrum. Near the emission maximum, F1 (∼ 0.35) is significantly smaller than at wavelengths near the edges of the profile (∼ 0.55) (see Figure S6 in Supporting Information). If the sample is excited with λex = 396.0 nm into the ligand ππ* manifold, F1 is reduced at all wavelengths but most significantly near the emission maximum (see Figure S6 in Supporting Information). This signifies that the second emitting state is likely ligand-based. The two emitting states could not be distinguished in transient absorption experiments because of the similarity of their lifetimes and the poor quantum yield of 3 in water. However, lifetimes of the emitting states are significantly reduced by the presence of oxygen in solution confirming that both states are triplet in nature. Temperature-Dependent Luminescence. The temperature dependence of emission quantum yields or lifetimes can provide valuable information about the relative energies of multiple excited states. Papanikolas et al. have used the temperature dependence of the quantum yield of [Ru(phen)2dppz]2þ to determine the relative energies of the Ru d-d excited state and the two 3MLCT excited states that contribute to its photophysical light switching behavior.37,38 Temperature dependence of radiative and non-radiative decay rate constants have also been used to probe the electronic effects of methyl substitution in substituted [Ru(phen)2(dpq)]2þ and [Ru(phen)2(dppz)]2þ complexes.105,106 Within this theme, the temperature dependence of the emission kinetics of 3 in various solvent systems has been measured (λex = 447.7 nm, λem = 600 nm) (Table 3). Predictably, the monoexponential emission lifetime of 3 in EtOH increases as the temperature is lowered because of reduced nonradiative decay. At 77 K, the emission decay is also monophasic with a lifetime of 4.2 μs. There are two possible explanations for the origin of the monophasic luminescence decay in EtOH at room temperature: the decay arises either from the isolated 3MLCT excited state or from a kinetic (105) O’Donoghue, K.; Penedo, J. C.; Kelly John, M.; Kruger Paul, E. Dalton Trans. 2005, 1123–1128. (106) Olofsson, J.; Wilhelmsson, L. M.; Lincoln, P. J. Am. Chem. Soc. 2004, 126, 15458–15465.

Figure 9. Triplet emission spectra in EtOH of 1 at 77 K (black), 3 at 77 K (red), 1 triplet sensitized with xanthone at room temperature (RT; blue), and 3 at RT (green).

steady state/equilibrium between the 3MLCT and the ligand localized 3ππ* state on a time scale faster than the luminescence.50 To distinguish between the two possibilities, uncoordinated ligand 1 was triplet sensitized at room temperature to estimate the energy of the 3ππ* state. Under this condition, strong phosphorescence from 3 is observed. Comparison of the luminescence of 3 (λex = 450 nm) with that of 1 sensitized with xanthone (λex = 355 nm, τ = 6.36 μs) reveals that at room temperature, the energy profile of the 3 MLCTphen state is nearly identical to the energy profile of the 3ππ*bptt state, with both displaying maxima at ∼600 nm (Figure 9). Since these two states are both triplet in nature and nearly degenerate, some degree of wave function mixing occurs leading to state coupling and the potential for rapid kinetic interconversion. The phosphorescence of 1 at 77 K also shows vibrational fine-structure with two distinct vibronic progressions (ΔE ∼1275 cm-1) corresponding to the energy of the C-C and C-N stretching modes of the aromatic rings.107 The vibrational signatures are consistent with those observed in the 77 K emission spectrum of 3 (ΔE ∼ 1225 cm-1), suggesting that the luminescence may be partially ligand in origin (Figure 9). The observation is also similar to other Ru(II) diimine complexes that display vibrational fine-structure with energies of 1050-1350 cm-1.66,96 To evaluate the luminescence of 3 in aqueous solution over a larger temperature range, a 60:40 ethylene glycol: H2O solvent mixture was chosen. As the temperature is raised from 233 to 333 K, the lifetime of the short-lived state increases while the lifetime of the long-lived state decreases. Concomitantly, the fraction of the emission from the short-lived state increases (Table 3), and is readily visible in the emission decay curves (Figure 10). At 333 K, the luminescence is nearly monoexponential with >99% of the emission arising from the short-lived 3 MLCTphen excited state. At room temperature, however, the luminescence trace is biphasic, illustrating that the rate of interconversion between the 3MLCTphen and 3ππ*bptt states is of the same order of magnitude as their luminescence decay (107) Campagna, S.; Puntoriero, F.; Nastasi, F.; Bergamini, G.; Balzani, V. Top. Curr. Chem. 2007, 280, 117–214.

11344 Inorganic Chemistry, Vol. 49, No. 24, 2010

Figure 10. Kinetic profiles of the excited state luminescence decay of 3 in 60:40 ethylene glycol:H2O at 233 K (black), 273 K (red), 298 K (blue), and 333 K (green).

rates. This further suggests that the two states are only moderately coupled with some appreciable inter-state barrier. At lower temperatures in aqueous solution, there is sufficient thermal energy for some fraction of molecules to traverse the barrier to the ligand-localized 3ππ*bptt state, but the higher kinetic barrier to the reverse rate retards backpopulation of the 3MLCTphen state. Consequently, the fractional emission from the 3ππ*bptt state (F2) is slightly larger at lower temperature (F2 = 0.07 at 298 K, F2 = 0.11 at 233 K). Thus, the lower energy 3ππ*bptt state effectively quenches the 3 MLCTphen state resulting in a reduced luminescence lifetime for the 3MLCTphen component of the emission. Meanwhile, the lifetime of the 3ππ*bptt state increases as the temperature is lowered because of reduced non-radiative deactivation. Interestingly, at 77 K, the emission is purely monophasic since excitation (λex=447.7 nm) leads to a thermally trapped 3 MLCTphen state that is no longer able to access the 3ππ* bptt excited state. The lifetime of 3 at 77 K in 60:40 ethylene glycol:H2O (τ = 5.0 μs) is similar to that of 3 (τ = 4.2 μs) and [Ru(bpy)2dppz]2þ (τ = 4.8 μs) at 77 K, but somewhat quenched relative to [Ru(phen)3]2þ (τ=9.9 μs) because of deactivation by the dark 3MLCTphzn state.38,66 In contrast, as the temperature is raised, the complex gains sufficient thermal energy to overcome the inter-state electronic barrier and the 3MLCTphen and 3ππ*bptt excited states increasingly equilibrate. At 333 K, the two states are almost completely equilibrated giving rise to the essentially monoexponential luminescence decay kinetics (τ = 143 ns). Excited State Structural Model. The excited state photophysics and dynamics model for 3 follows from the established electronic paradigms for both [Ru(phen)2dppz]2þ31,39-41,45,46,108 and [Re(dppz)(CO)3L] analogues51,52 where two 3MLCT states (phenathroline- and phenazine-based) and a 3ππ* state control the observed light switch behavior and photophysics. From the temperature dependence of [Ru(phen)2dppz]2þ emission in solvents of different polarity, Meyer et al. have shown that the lowest lying 3MLCTphzn state is charge-separated in nature, while the 3 ππ* state of dppz lies to higher energy (∼540 nm; τ ∼ 10s μs) and does not contribute to the observed luminescence.38 (108) Coates, C. G.; Jacquet, L.; McGarvey, J. J.; Bell, S. E. J.; Al-Obaidi, A. H. R.; Kelly, J. M. J. Am. Chem. Soc. 1997, 119, 7130–7136.

Spencer et al.

Figure 11. Relative energy level diagram of the interactions between the phenanthroline-based 3MLCT excited state and the ligand-based 3π-π* excited state for 3 in ethanol (a) and water (b).

In contrast, fac-[Re(dppz)(CO)3Cl] exhibits 3ππ* ligandbased luminescence at 77 K and 3MLCT luminescence at room temperature, and chloride substitution with triphenylphosphine leads exclusively to luminescence from the 3ππ* state at both 77 K and room temperature.52 Resonance Raman studies have shown that the 3ππ* state is lower in energy for both complexes, but in the case of fac-[Re(dppz)(CO3)Cl] the two states are nearly degenerate such that the 3 MLCT state is thermally back-populated at room temperature. Since the radiative and non-radiative kinetics of the 3 MLCT are much faster than the ligand-based triplet state, the emissive 3MLCT state dominates the observed luminescence as predicted by inter-state conversion kinetics.50 Additionally, picosecond time-resolved infrared spectroscopy shows that the relative population of the two states in fac-[Re(dppz)(CO3)Cl] is solvent dependent, and if electron-withdrawing substituents are added to the dppz ligand, population of the 3ππ* state is no longer observed because of lowering of the dark state energy.51 The photophysics of 3 are encompassed by this general model. Three excited states lead to the observed luminescence behavior, two 3MLCT states and the 3ππ*bptt ligandbased state that lies to lower energy and plays an important role in the observed photophysical behavior. From emission and triplet-sensitized phosphorescence measurements of 1, the 3MLCTphen and 3ππ*bptt states of 3 are of similar energies (λ ∼ 600 nm), but the 3MLCTbptt state is significantly lower in energy (λ > 800 nm) and exhibits no detectable emission. Consequently, it is the interactions between the 3MLCTphen and 3ππ*bptt states that give rise to the unique solventand temperature-dependent luminescence behavior of 3 (Figure 11). Chelation of the bptt ligand to Ru(II) shifts the 3ππ*bptt state slightly to lower energy (λem ∼ 620 nm) relative to the free ligand upon triplet sensitization (λmax = 600 nm). When solvated in ethanol at room temperature, the 3 MLCTphen and 3ππ*bptt states are strongly coupled, and the small barrier between them leads to rapid interconversion (Figure 11a). The observed emission maximum (λem = 603 nm) and monoexponential lifetime (τ = 630 ns) of 3 indicate that the primary luminescence derives from the 3MLCTphen state that is perturbed by the presence of the 3ππ*bptt state close in energy. This is consistent with the fact that the 3 MLCTphen state has a lifetime that is approximately an order of magnitude shorter than that of the 3ππ*bptt state (τ = 6.6 μs at room temperature), and suggests that the

Article

Inorganic Chemistry, Vol. 49, No. 24, 2010

11345

thermal trapping in the 3MLCTphen state occurs leading to monoexponential luminescence kinetics.

rate of interconversion between the two states is more rapid than decay to ground state. In an aqueous environment, the 3MLCTphen and the 3 ππ*bptt states are both stabilized relative to their energies in ethanol. The charge-separated 3MLCTphen excited state experiences a larger stabilization of ∼800 cm-1 that is consistent with that observed for the 3MLCTphen excited state of [Ru(phen)2dppz]2þ.9,45,47 The dielectric stabilization reduces the energy gap between the 3MLCTphen and 3ππ*bptt states, and combined with the enhanced reorganizational energy in water, increases the activation barrier between the two states. The increased thermal barrier does not allow rapid interconversion at room temperature (Figure 11b). As the rate of interconversion slows and approaches the rate of radiative decay, emission from both excited states results. The energies of the 3MLCTphen and 3ππ*bptt states correspond to approximately 610 and 620 nm, respectively, when bound to DNA (λem = 615 nm) based on the measured shift in the time-resolved luminescence making the energy gap between the two states ∼260 cm-1. If the energy gap between the states is approximately the same in water (λem = 634 nm), the two states would correspond to emission wavelengths of ∼630 and 640 nm. In a complementary approach, the energy gap can be estimated from the temperature dependence of the luminescence by using the Boltzmann distribution, and approximating the population of each state by the observed luminescence intensity.109-112 In this case, since only one emission band is observed, the fractional components of the emission lifetimes were used as an estimate of the populations. A plot of ln(F1/F2) versus 1/T was constructed, and the slope was used to get a rough estimate of the energy gap. The energy gap was found to be ∼350 ( 100 cm-1 in 60:40 ethylene glycol/H2O, which is similar to the estimate from the time-resolved emission data of 3 bound to DNA (see Supporting Information, Figure S7). The dual emission behavior suggests a higher kinetic barrier in H2O relative to ethanol is due to a larger solvent reorganizational energy during conversion from the polar, charge-separated 3MLCT state to the ligandlocalized 3ππ*bptt state. This is consistent with Ru(II) diimine complexes that show deviations from the energy gap law in protic solvents, and in particular H2O, because of the large reorganizational energy.113-115 At 77 K in aqueous solution, the rate of interconversion between the two states is slow in part because the high barrier and

The photophysical properties of 3 and 4 in ethanol resemble those of [Ru(phen)3]2þ, indicating that traditional MLCT emission is operative and only sparingly affected by the extended conjugation of the bptt framework. In an aqueous environment, 3 and 4 possess notably small luminescence quantum yields consistent with non-radiative solvent quenching at the phenazine nitrogen lone pairs. Moreover, addition of CT-DNA to an aqueous solution of 3 causes a significant increase in the luminescence quantum yield while the quantum yield of 4 is relatively unaffected indicating that tert-butyl substitution on the terminal phenyl groups inhibits the ability of 4 to intercalate with DNA. In contrast to that of [Ru(phen)2dppz]2þ, the solvent- and temperature-dependence of the luminescence of 3 reveal that the extended ligand aromaticity lowers the energy of the 3ππ* excited state into competition with the emitting 3MLCT state. Interconversion between these two states plays a significant role in the observed photophysics and accounts for the dual emission in aqueous environments. These results have key implications to the development of metallo-probes of DNA structure to further enhance our understanding of the subtleties of minor groove association, intercalation, and insertion.7 As more advanced probes with more specialized steric considerations or further extended πstructures become prevalent, 3ππ* ligand localized states can mix and compete with the 3MLCT manifold to perturb the excited state dynamics of these probes and lengthen their lifetimes.50 In some cases, this may not be pragmatically important as rapid equilibrium conditions relative to radiative decay or excited state trapping because of markedly slower equilibration rates with respect to luminescence to the ground state may occur. These conditions can lead to straightforward monoexponential decays that are somewhat invariant across a biological experiment. However, these same rate constants are also solvent environment sensitive, and because of thermodynamic and reorganizational energy considerations, can vary as a function of parameters such as dielectric/cosolvent, temperature, and potentially ionic strength. Therefore it is critical to carefully understand the specific excited state probe dynamics to appropriately interpret binding data from one DNA binding experiment to the next.

(109) Wang, Z.; Lees, A. J. Inorg. Chem. 1993, 32, 1493–1501. (110) Zulu, M. M.; Lees, A. J. Inorg. Chem. 1989, 28, 85–89. (111) Kirchhoff, J. R.; Gamache, R. E., Jr.; Blaskie, M. W.; Del Paggio, A. A.; Lengel, R. K.; McMillin, D. R. Inorg. Chem. 1983, 22, 2380–2384. (112) Watts, R. J. Inorg. Chem. 1981, 20, 2302–2306. (113) Sun, H.; Hoffman, M. Z. J. Phys. Chem. 1993, 97, 11956–11959. (114) Caspar, J. V.; Sullivan, B. P.; Kober, E. M.; Meyer, T. J. Chem. Phys. Lett. 1982, 91, 91–95. (115) Caspar, J. V.; Meyer, T. J. J. Am. Chem. Soc. 1983, 105, 5583–5590.

Supporting Information Available: X-ray crystallographic data of 3 in CIF format, comparisons of X-ray structure and DFT calculated structure for 3, resonance Raman spectra of 3 and 4, experimental Raman frequencies and calculated normal modes for 3 and 4, DFT molecular orbitals of 1, TDDFT spinallowed electronic transitions of 1, plot of F1 across the emission spectrum, and Boltzmann plot. This material is available free of charge via the Internet at http://pubs.acs.org.

Summary and Implications